The AI-Human collaboration behind our content (Part 1)
Venture capital firms see countless insights daily: patterns spotted in portfolio reviews, interesting conversations with founders, regulatory changes, macro trends, and academic research. Today, we're pulling back the curtain to show you exactly how we communicate and build on ideas at Good Capital. We’ll illustrate how we inculcate AI applications in practice and show you how this newsletter is created.
The challenge isn't generating ideas; it's capturing, synthesising, and channeling them into coherent communication that serves different audiences.
Most teams struggle with this coordination layer. We try to treat this as a workflow design problem, not necessarily a content problem. Our approach solves three critical challenges most content teams face:
Preserving context as ideas move from meetings to public content
Enabling the same insight to be adapted for different audiences without losing coherence, and
Preserving quality standards across all team members and channels.
The whole system works because it amplifies what we're already good at rather than trying to replace it. We still do all the thinking and creating - we just don't get bogged down in the coordination anymore.
The idea generation ecosystem
We begin each quarter by discussing the 3-4 key ideas that we feel strongly about and want to communicate. These quarterly narratives become the strategic framework that guides every piece of content we create.
Within this framework, our content ideas emerge from five distinct human activities:
Small Ideas Meetings: Weekly sessions where each team member pitches the 5 most interesting things they came across: articles, founder conversations, and portfolio company tactics.
Big Ideas Meetings: Monthly sessions where the entire team reads identical source material - academic papers, podcast series, industry reports, or books - then debates implications for our investment strategy and thesis.
Investment Reviews: Weekly investment review meetings where we discuss portfolio performance and pipeline companies.
Development Meetings: Bi-weekly marketing sessions where we try to develop insights from the above activities into compelling ideas.
The AI foundation
On the distribution side, we have dedicated Claude projects for each of our audiences, one each for LPs, one for founder-specific communications, and one for social media. Each project has become a specialised AI collaborator with a deep understanding of its specific audience & format requirements.
Each project has the data of:
Custom Instructions: Detailed project knowledge that defines the project's role, audience, and objectives, ensuring Claude understands exactly what we're trying to achieve for each channel.
Narrative Resources: Documents that brief Claude on our current quarterly themes and how to weave them naturally into content, ensuring every piece aligns with our strategic messaging priorities.
Style Guide Resources: Our complete archive of previous content for that channel - newsletters, blog posts, and Twitter threads - that function as an implicit style guide, teaching Claude our voice, tone, and format preferences through examples rather than rules.
Positioning Resources: Materials that give Claude context on who we are as Good Capital, how we position ourselves differently within the VC ecosystem, and our unique perspective on Indian markets and AI-powered intermediaries.
Monthly Updates - Each project gets refreshed monthly with new content, performance feedback, and audience insights, creating AI collaborators that continuously learn and improve.
This systematic foundation - quarterly narratives, ongoing idea generation, and specialised AI infrastructure - creates the backbone for everything we publish. But having great ideas and smart AI systems is only half the story.
In the next part, we'll show you how these captured ideas and AI collaborators work together in practice: how we evaluate which ideas fit which audiences, and the actual back-and-forth process that turns meeting discussions, insights, and research into the newsletters you read.